By Shabnam Wazed, Founder and CEO of AGAM International

Banks and lenders have historically focused extensively on a borrower’s current financial information when evaluating their likelihood of repaying a loan on time. Lending decisions are made on factors like payment history and outstanding debt. But millions of conscientious people - potential borrowers who would move mountains to repay their loans, remain without the resources they need.

In many countries, including the US, a borrower’s creditworthiness is represented by a three-digit numerical score. If the borrower’s score is above the lender’s threshold, the loan is approved; if the score is below, the application is rejected. The score is dependent on the number of transaction histories and the higher the volume the higher the score. Therefore, the likelihood of loan repayment is reliant on how many transactions and payments the customer has had in the past, rather than on if they are deemed suitable to repay it in the future.

In recent years, concerns about the effectiveness of traditional credit scoring models have begun to surface. These concerns, which include the inability to serve countless millions of “unbanked” and “underbanked” consumers around the world, have led fintechs and financial institutions to explore non-traditional credit score models.

This shift coincided with the emergence of Big Data which, in combination with advanced behavioural analytics, has given fintechs the leverage to develop accurate alternative credit score models to supplement the more traditional transaction-based approach.

So is now the time switch how we score – for both consumers and financial institutions?

Below are the five drivers which indicate that now is the time to evolve the model.

1)    Prevalence of Big Data usage

Oracle defines Big Data as “data that contains greater variety, arriving in increasing volumes and with more velocity.” From a more practical standpoint, it is the enormous volume of data people generate every day that leaves behind a trail of information that can be used not only to learn about our preferences, but to accurately predict our future habits.

From scanning a loyalty card at the supermarket to choosing a movie or series on a favorite streaming service, our choices join the flood of Big Data.

Although a lot of information about customers is held, studies have shown these enormous datasets are too diluted to create a conclusive picture about an individual’s borrowing behaviour. In short, Big Data alone lacks the ability to predict financial behaviour of customers and falls short of making sophisticated risk modelling to generate an alternative credit score (Biatat et al, 2021).

However, the dataset is key for fintechs setting up in this space. The data can be augmented with behavioural analytics with a deeper focus into an individual’s responsiveness, learning behaviour, mental and emotional state, tendencies, habits, and beliefs. Giving a multi-tiered insight into the customer, their character and their context can lead to a high predictive accuracy towards the likelihood of repaying debt (Klinger, Khwaja and LaMonte, 2013)

2)     Why are behavioural analytics important?

The use of behavioral analytics and psychometrics in alternative credit score models is significantly shifting the way people can borrow. Behavioral analytics also gives an opportunity to a borrower to prove their genuine intent to repay, and allows banks and financial institutions to predict defaults  and make better informed lending decisions by taking into account a range of factors reflected from the customer.

3)    Using psychometrics to fill the gaps

Psychometrics can help potential borrowers whose credit histories are incomplete. This is particularly true in emerging markets where credit bureaus may not be subject to national regulations, or where technology or other limitations have led to fractured reporting and communication.

Access to finance remains a challenge for many micro, small and medium-sized enterprises (MSMEs). Lenders lack the tools to reach these borrowers with sufficient scale and control over risk due, in part, to the shortcomings of screening technologies. In 2012 a pilot of an innovative psychometric tool aimed at evaluating credit risk for business owners seeking a loan was conducted in Peru, where MSME loans were disbursed by Financiera Confianza, for merchants with limited transaction history. The goal was to determine if, and to what degree, this tool could assess an enterprise’s level of risk to the lender. The results suggested a powerful contribution to risk analysis that was highly scalable and could be implemented, even on unbanked business owners with no credit history or collateral.

4)    Helping the “financially invisible”

For individuals and small businesses with no credit or banking history – sometimes referred to as the “financially invisible” - psychometrics can give them a means of qualifying for a loan. These potential borrowers typically face the chicken and egg conundrum of needing a loan to establish credit history, but being unable to provide a history because of the inability to obtain that first loan.

These consumers make up a vast potential audience for fintechs, with the vision and flexibility to pursue the global market. In many emerging markets, alternative credit scoring through AI and question-based psychometric scoring, could help significant portions of the population thrive.

Globally, 1.6 billion adults are classified as unbanked - that is, they do not have access to a bank or other financial organisation.

Morocco tops the list of countries with the largest percentage of unbanked adult citizens at 71 per cent, with Vietnam, Egypt, the Philippines, and Mexico rounding out the top five. Together, they represent over 300 million consumers with currently no ability to obtain traditional banking services or credit.

Alternative credit scoring models involving behavioural analytics provide the opportunity for these borrowers to build their credit histories to provide better futures.

5)    Improving credit decision making

Obtaining and evaluating psychometric information can also provide insights beyond the individual’s financial ability to repay. Although ability is a critical factor, it’s equally important to consider an applicant’s willingness to repay their loans on time.

Why? Refusal to pay is among the top reasons for loan default.

In some cases, borrowers explicitly refuse to pay based on one or more stated justifications. In many other cases it may be a person’s subconscious beliefs or habits that cause borrowers to de-prioritise and even default on their loans.

The use of psychometrics aims to cultivate insights taken from Big Data to infer an applicant’s behavioral traits, attitude toward money and debt, and other factors that will substantiate their repayment potential.

This allows them to not only spot potential warning signs such as a high debt-to-income ratio or a recent history of missed payments, but other signs such as tendencies toward excessive spending or unchecked impulse shopping, negative attitudes toward debt repayment, or social influences that may affect an applicant’s credit behavior.

Such analytics-based alternative credit scoring offers lenders superior insights to help them reduce their overall portfolio risk by more accurately identifying and approving borrowers who have the highest likelihood of repayment. By choosing more responsible borrowers, they reduce enormous collection costs, losses from loan defaults, and other expenses caused by inaccurate lending decisions.

The way forward

These dynamics open new possibilities for fintechs and lenders to increase the size of their portfolio.

There is fierce competition among lenders who want to profit from well-managed lending. Borrowers with excellent credit scores are constantly inundated with offers for more credit - more offers than they could possibly accept.

In the meantime, millions of conscientious people - potential borrowers who would move mountains to repay their loans, remain without the resources they need.

By focusing on emerging markets and providing lending and financial services to the unbanked, they can strategically use these models to identify excellent-fit borrowers and create long-term lending relationships.

AI-based credit scoring can help lenders succeed and should be weighted at least equally with traditional credit scoring models, if not more heavily.

Alternative credit scoring gives lenders and borrowers the opportunity to grow, instead of reserving these opportunities for people who already have access to money and have a choice to borrow or not. The global growth of fintechs, working with banks and lenders, means those people who have not traditionally been granted access to formal loans can now finally stop borrowing from high interest loan sharks and be able to plan for their futures. It is a win-win situation for all involved.